Strategies of distributed genetic algorithms for three-dimensional bin packing in a SLS machine

This dissertation details the creation of three distributed genetic algorithms used for bin packing in a Selective Laser Sintering (SLS) machine that is used for rapid prototyping. The three different distributed genetic algorithms were implemented: a distributed chromosome, a multiple copy server, and a hybrid.; This dissertation demonstrates that using a distributed chromosome implementation reduces the execution time of a genetic algorithm. It also demonstrates that using a hybrid distributed implementation or a multiple copy server implementation produces better fitness values. Both were statistically tested and verified with an experimental design.; The space and time complexity was also calculated for the sequential genetic algorithm as well as each of the three distributed genetic algorithm implementations. Several test cases were executed to validate the theoretical time and space complexities with different number of parts packed in the SLS build cylinder.; In addition, this dissertation also demonstrates that a hybrid distributed genetic algorithm or a multiple copy server genetic algorithm has less dependence on the value of the probability of crossover and the probability of mutation. This is due to the fact that the hybrid or multiple copy genetic algorithms have the ability to use multiple probability values simultaneously.